Ask Your Question

crackwitz's profile - activity

2019-10-04 17:23:52 -0500 answered a question self intersected contour

contours are polygons. you'll find more material when you search for that instead.

2019-10-04 17:18:11 -0500 answered a question Is there an error in the color conversion components?

I see that your Green value is 0. that could be clipping. I would say you lost information right there. try it with dif

2019-06-26 11:50:30 -0500 commented question How can I use Hough Transforms to fit a Plane

you have point clouds? then hough transform sounds useless. I would recommend PCA, as mentioned above, or RANSAC.

2019-06-26 11:43:13 -0500 commented question how to calibrate two cameras placed up and down?

what does "up and down" mean?

2019-06-26 11:40:44 -0500 answered a question Exception thrown when converting OpenCV Mat to array for 1024x1024 but not 512x512?<double>(i, j); that is the problem you have. check the type of mat0. it is probably 8-bit unsigned integ

2019-06-26 11:33:13 -0500 answered a question Direct conversion of pixels to millimeters?

use intrinsic calibration parameters to turn the pixel coordinate (in 2d screen space) into a viewing ray (in 3d camera

2019-03-04 22:00:12 -0500 answered a question How to connect Nikon Coolpix P900 camera ?

if you want to capture HDMI, you'll need a HDMI capture device for this. when you search for "HDMI capture USB3" you wil

2019-02-20 19:14:48 -0500 answered a question Stereo Camera Calibration Problems

tl;dr: hold the board very still or get a camera with global shutter. Rolling shutter, a property of most webcams and s

2018-06-01 13:45:07 -0500 received badge  Enthusiast
2018-04-25 14:25:27 -0500 commented question High cpu usage in OpenCV java

decoding 4k streams costs cpu time. figure out how to decode with hardware support. what's available depends on your har

2018-04-25 12:31:55 -0500 commented answer Professional image recognition help needed

does this involve robotics? you're looking at some real expenses for R&D here, paying human and buying the proper eq

2018-04-25 11:04:05 -0500 answered a question Professional image recognition help needed

verrrry hard if those are images one would have to work with. in any case, I would refuse to work with these images and

2018-03-17 14:52:47 -0500 commented question Convert Image to Occupancy Grid

your scene is synthetic, your camera and lighting looks static. if you have an image of the "background", i.e. no obstac

2018-03-17 14:48:59 -0500 answered a question Compare two images and mark the unmatched object or remove unmatched

what do you have so far?

2018-03-17 14:48:15 -0500 answered a question Detection of wounds

I agree, hard to impossible, as is. what is the problem you're trying to solve with this?

2018-02-22 09:14:32 -0500 received badge  Popular Question (source)
2014-04-19 19:49:11 -0500 answered a question writing MJPG file by VideoWriter in Visual Studio x64/Release platform

it's probably a floating point comparison issue. you do not compare floats for equality. you compute the absolute or relative error and check if it's below some acceptable threshold "epsilon".

alternatively: something happens to EXPECT_EQ(). you'll need to investigate, what this function (macro?) does in a release build, or an x64 build.

2014-04-19 19:47:39 -0500 answered a question videowriter using ffmpeg h264 codec on windows with opencv 248

I haven't gotten it to work properly either. I say "fourcc = -1" and then use the system codec picker to choose x264vfw. if I give "X264" as the codec, it complains about "codec ID 28" but still produces a video (provided I ignore isOpened() being false). when I give a completely random fourcc, it doesn't complain about "codec id 28" and doesn't produce a video.

if you give a fourcc, it's passed to avcodec_find_encoder() in sources\modules\highgui\src\cap_ffmpeg_impl.hpp line 1591 (in v2.4.8), so somewhere there is the bug.

2014-03-14 10:14:58 -0500 received badge  Editor (source)
2014-03-14 10:05:25 -0500 asked a question VideoReader_GPU not available, but built with NVCUVID?

I've built OpenCV 2.4.8 with CUDA 5.5 using Visual studio 2012. Building code to use it seems to work too. gpu::getCudaEnabledDeviceCount() reports 1.

However, when I try to initialize a VideoReader_GPU, I get this runtime error:

OpenCV Error: The function/feature is not implemented (The called functionality is disabled for current build or platfom) in throw_nogpu, file P:/opencv-2.4.8/sources/modules/gpu/src/precomp.hpp, line 131

CMake reported availability and successful build of NVCUVID. How can I verify that? What's going on?

EDIT: I've just perused the sources I downloaded... #if !defined(HAVE_CUDA) || defined(CUDA_DISABLER) || !defined(HAVE_NVCUVID) should be responsible. I will try to figure out which define is causing this...

EDIT: I've changed that line to remove CUDA_DISABLER. that seems to have worked... since both HAVE_NVCUVID and HAVE_CUDA are reported to be true/defined by CMake.